[BioC] help in extracting limma results

[echang4@life.uiuc.edu]

Hi Bioconductor users,
I am having trouble understanding how multiple-testing adjustment is done
in limma (specifically the decideTest).  I am really confused about the
meaning between moderated F p-value and the adjusted p-value.

If I try two different "flavors" of decideTest (e.g. nestedF and global),
I can see that the results are different.... but is it the p-value that is
adjusted or the F-statistic is adjusted?
And how do I extract the gene lists (cutoff of FDR-adjusted p-value< 0.05)?

I basically followed the limma user guide (which is superbly written,
thank you!) But after completing the analysis, I am stuck trying to
extract the gene lists out of R and into a csv file.

Can someone please help me out with the code which would extract the genes
under "results2"? My code is as follows:

> cont.matrix <- makeContrasts(
Ad.V_E2 = Ad.E2-Ad.veh,
b20.V_E2 = b20.E2-b20.veh,
Veh_modul = b20.veh-Ad.veh,
E2_intrxn = (b20.E2-b20.veh)-(Ad.E2-Ad.veh), levels=design)
>fit2 <- contrasts.fit(fit, cont.matrix)
>fit2 <- eBayes(fit2)
>results2 <- decideTests(fit2, method="nestedF")
>summary(results2)
   Ad.V_E2 b20.V_E2 Veh_modul E2_intrxn
-1     729      699      1108       224
0    20651    20662     20662     21830
1      903      922       513       229

> results2 <- decideTests(fit2, method="global")
> summary(results2)
   Ad.V_E2 b20.V_E2 Veh_modul E2_intrxn
-1     501      460       641       101
0    21174    21166     21311     22064
1      608      657       331       118
>write.csv(topTable(fit2, coef="Ad.V_E2", n=22283), "Ad+E2.csv",
row.names=FALSE)
***If I do this, won't I end up with the unadjusted p-values?***

I also see other examples like this....how can I export genes in
"selected"?
> selected<- p.adjust(fit2$F.p.value, method="fdr") < 0.05

Basically, I'm stuck.
Any help would be greatly appreciated. I really did read the limma user
guide, but if anyone could point sections which would help me, I'd very
much appreciate that as well.

Thank you,
Edmund Chang